I'm working on a home navigation robot. It will share a few things with the DB1 such as using mostly Actobotics for the chassis and I'm sure some of the design goals will overlap as well. It will be using an RPI for the main computer along with an Atmega based board (a nano for now for the motors/encoders) and one Pic based board (an Uno32 Pic chip for the sensors) I am using a couple of custom boards I have done for the Nano and one for an I2C bus. It will use a Pololu Qik serial motor controller to drive the motors via the nano.
My design goals now are:
- Medium sized, about 1.5ft (46cm) cubed weighing less than 6.5lbs (3kg)
- Have a mapping system using premade maps and ad-hoc additions to maps by the bot.
- Limited voice control.
- Wifi/Web/REST control/monitoring.
- Modest sensor array (Sonar, IR, and good ole bumper switches)
- Limited Vision navigation via Pixy cam to start and later OpenCV or OpenMV. This will include some bot handled object detection.
- Room for additions and experiments.
Find the latest on it at: https://randomcoderblog.wordpress.com/tag/robotics/
Scott
I took a look at Nav Bot blog. You have a really nice robot project there. It's nice to see it well documented.
DroneBot Workshop Robotics Engineer
James
Thanks. It's also a good way to think out loud.
Great project, thanks for sharing.
Thanks. I am hoping to get some work done on it this weekend. Work and home life gets in the way 😉
Did a tracking test with no encoders being read. I want the bot to drive as straight as possible with the encoders only doing touch-up work. The turns are about as consistent, but the omni-wheel does change it as you move.
The plush carpet serves as a nice recorder of the tracks. Looks to be quite dependable in terms of precision repeatability. Looks like you did a good job getting the wheels synchronized.
DroneBot Workshop Robotics Engineer
James
Leaving lines over the carpet doesn't go down well with the wife!
My suggestion is you run it along a long passage because any slight turning may not be noticeable over a short distance and reversing just reverses that slight turn?
My test rig uses two iRobot motors with hall effect encoders and I was surprised to find that one motor runs at only 78% the speed of the other motor when the same pwm is applied to both at the same time.
My suggestion is you run it along a long passage because any slight turning may not be noticeable over a short distance and reversing just reverses that slight turn?
This is true. Going back and forth over a short distance could actually be a very slight arc that is simply consistent. At least it still shows consistency. But it may not be going straight.
DroneBot Workshop Robotics Engineer
James
My suggestion is you run it along a long passage because any slight turning may not be noticeable over a short distance and reversing just reverses that slight turn?
My test rig uses two iRobot motors with hall effect encoders and I was surprised to find that one motor runs at only 78% the speed of the other motor when the same pwm is applied to both at the same time.
I have found you usually get a large pull in starts, due to the motors clamoring for current (stupid electricity). I'm starting with a ramping function up and down to keep this to a minimum. I will be doing longer tests, but am in my office now and without going out in the wilds of my home, short will do for now. I have done several minutes of that and in turns. The turns do drift due to the omni wheel bumping along at different angles. I do have encoders though and will correct with them.
On your iRobot, are the motors match in polarity? I mean, do you have them both able to go in the same direction when you go forward or reverse? Some motors go slower in reverse than forwards. I find having them matched this way helps keep them better matched with both either going forward or in reverse near the same RPM. Another thing to watch for is having your PWM IO lines running at the same frequency. This can cause drift as well if they are on diffent timer freqs.
Scott
Great job jscottbee. Hope you complete it soon.
nice bot project
Nawar Alseelawi
A very small update to my project.
Great project. For some reason the link on the last update does not work. The link on the June 15th post worked just fine.
Thanks. I goofed on the link copy 🙂